ML13330B561
| ML13330B561 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 09/20/1991 |
| From: | Rosenblum R SOUTHERN CALIFORNIA EDISON CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| NUDOCS 9109270116 | |
| Download: ML13330B561 (12) | |
Text
Southern California Edison Company 23 PARKER STREET IRVINE, CALIFORNIA 92718 R. M. ROSENBLUM TELEPHONE MANAGER OF (714) 454-4505 NUCLEAR REGULATORY AFFAIRS September 20, 1991 U. S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D.C. 20555 Gentlemen:
Subject:
Docket No. 50-206 Acceptance Criteria for the Thermal Shield Monitoring Program San Onofre Nuclear Generating Station Unit 1 The enclosure to this letter provides the acceptance criteria for the reactor vessel thermal shield monitoring program in accordance with License Condition 3.M, "Cycle 11 Thermal Shield Monitoring Program" of Provisional Operating License DPR-13. This submittal satisfies open items (e) and (f) from the NRC Safety Evaluation Report dated December 19, 1990, which accepts the design of the thermal shield supports.
The thermal shield monitoring program consists of neutron noise monitoring and loose parts monitoring. The acceptance criteria for both are included in the enclosure to this letter. The acceptance criteria for neutron noise were established following baseline evaluations performed at 85% power or greater for 60 days as required by License Condition 3.M.(1).
The acceptance criteria for the loose parts monitoring system were determined after monitoring the background noise at a variety of power levels and performing calibration tests during the beginning of Cycle 11 operation.
The acceptance criteria in the enclosures will be used to operate San Onofre Unit 1 in accordance with License Condition 3.M during Cycle 11 operation.
The acceptance criteria for the thermal shield monitoring program will later be incorporated into a new proposed technical specification for plant operation during Cycle 12 and beyond. In accordance with open item (c) from your December 19, 1990 letter approving the thermal shield replacement support design, this new specification will contain the requirements for monitoring FDR' pIC 6500U0206
Document Control Desk
-2 the condition of the thermal shield during plant operation and for inspecting the thermal shield support system during periods when the plant is shutdown.
We will provide this new proposed technical specification by January 31, 1992.
If you have any questions, please do not hesitate to call.
Very truly yours, Enclosure cc:
J. B. Martin, Regional Administrator, NRC Region V George Kalman, NRC Senior Project Manager, San Onofre Unit 1 J. 0. Bradfute, NRC Project Manager, San Onofre Unit 1 C. W. Caldwell, NRC Senior Resident Inspector, San Onofre Units 1, 2&3
THERMAL SHIELD MONITORING FINAL ACCEPTANCE CRITERIA Monitoring of the San Onofre Unit 1 reactor vessel thermal shield is accomplished by the Neutron Noise Monitoring and Loose Parts Monitoring Systems. Two separate methods are used to ensure independent corroboration of any trends in thermal shield condition in the unlikely event of thermal shield degradation.
Neutron Noise Monitoring Methodology The ex-core detector electrical current signals are fed directly to a personal computer for analysis.
All eight neutron noise inputs are monitored for at least twenty minutes once each week.
At least 3 input signals are required to be functional for monitoring, in accordance with License Condition 3.M, Section 3.b. The signal data is processed by the PC software by Power Spectral Density (PSD) and Cross Power Spectral Density (CPSD) analysis, and the analysis includes consideration of coherence and phase between detectors in developing thermal shield frequency and amplitude data for comparison with baseline information. Both in-phase and out-of-phase PSDs are processed to enhance the signal analysis.
The processed data are reviewed weekly for compliance with acceptance criteria. The data are also transmitted to Westinghouse via telephone modem to add to the Westinghouse database and for further review and evaluation when warranted. Floppy disks containing the data and hard copies of the plots are also mailed as a back up, as necessary.
Neutron Noise Final Acceptance Criteria Acceptance criteria for Cycle 11 were developed based on baseline information obtained between April 2, 1991 and July 15, 1991 over a reactor power range of 88% to 92% power. Eighteen separate data sets were included in this baseline information. In addition, due consideration was given also to historical data obtained during the entire Cycle 10 operating period, including the.baseline information used in developing Cycle 10 criteria. The data analyses are documented in Westinghouse report WCAP 13058 (Ref. 5), and final acceptance criteria were established based on cumulative increases or decreases in neutron noise PSD levels and changes in the center frequency of spectral peaks. In addition, correction factors were developed to account for fuel burn-up effects on amplitude levels over the Cycle 11 operating period.
Baseline envelopes were established for the analyzed data sets for each of the eight ex-core detectors, using the Cycles 10 and 11 baseline data sets. Using the envelopes, curves were established for each detector at 200% above the corresponding envelope upper bound curve, and similarly at 33% of the lower bound curve. For the 6 Hz and 13 Hz peaks, the upper and lower criteria bounds were widened to 300% and 25%, respectively, of the Cycle 11 envelopes.
Curves were also developed at 125% of the upper bound and 44% of the lower bound curves for these peaks. The curves also include changes in the spectral
peak frequencies to properly account for expected frequency shifts that could be attributed to thermal shield degradation. These changes provide additional margin in the criteria for avoidance of false alarms while still remaining conservatively within the amplitudes expected from degradation of the thermal shield. The curves are given in Figures 1 through 4.
The inner curves establish initial review boundaries. Noise level data outside of these boundaries, after appropriate adjustment for burn-up effects, requires analysis and evaluation of data for trends.
The outer curves establish the boundaries committed to in License Condition 3.M. Noise level data outside of these boundaries, after adjustment for burnup effects, requires action in accordance with the License Condition, including notification of the NRC.
The boundary curves provided here have a wider range than those developed during Cycle 10.
This is justified based on the availability of additional data from Cycle 10 and Cycle 11, and on Westinghouse's additional work done validating the analytical model of the thermal shield/core barrel system.
This latter work is documented in WCAPs 12981 (Ref. 6) and 13047 (Ref. 7).
Nevertheless, the upper bound amplitude levels are still below the levels predicted for thermal shield shell modes for postulated thermal shield worst credible degraded case conditions. It should also be noted that the curves are developed only for the frequency range of interest: 1 Hz to 15 Hz.
Increases in neutron noise signal levels have been attributed to core burn-up and boron concentration changes in many plants. As a larger neutron noise data base is accumulated, the trends that are caused by changes in core burn-up and/or boron concentration are being substantiated. Increase in neutron noise levels due to burn-up and boron depletion are not uniform for all plant designs. Reference 1 notes that these effects can differ significantly depending on the fuel loading or design. Burn-up effects should not cause changes in natural frequencies attributable to thermal shield and core barrel motions, although a shift in fuel assembly natural frequencies over the first fuel cycle was inferred from the data reported in References 1 and 2. Reference 5 provides for evaluation and quantification of burn-up effects for SONGS 1.
When the inner curve levels are reached, the SONGS 1 data will be evaluated.
Interpretation of the data will be supported by comparison of the characteristics of the trends that have been observed in the data of other plants (examples are provided in References, 1, 3, and 4).
As reported in Reference 1, an essentially linear increase in RMS signal level occurred with "soluble boron concentration changes associated with fuel burn-up." In addition, no change in the center frequency of peaks is expected to accompany trends due to burn-up and boron depletion effects. Westinghouse experience is consistent with the above.
The trends established for the Unit 1 data are summarized in Figure 5, which provides correction factors for amplitudes based on historical information for Cycle 10 and the first 60 EFPD of Cycle 11 for Unit 1. These correction factors will be applied to the signal data whenever levels outside the bounds of the criteria curves are obtained. If the corrected data does not exceed
the criteria envelope after correction, the data will be considered as being within the criteria, and no further action will be taken.
The correction will be applied as follows:
PSDe = PSDd / [1 + (CF) (EFPD)]
2 where PSDe
=
Power Spectral Density Amplitude (PSD) at frequency exceeding the guidelines, adjusted.
PSDd=
PSD at frequency exceeding guidelines, to be adjusted.
=
Correction Factor at frequency that PSD exceeds guideline criteria, from Figure 5 EFPD
=
Number of effective full power days of core burn-up from start of Cycle 11.
The following actions will be taken, depending on the amplitudes of the data collected during ongoing neutron noise monitoring:
- 1.
If the inner "evaluation curve" levels are reached at any frequency within the criteria, detailed examination of the trends of amplitudes and center frequencies of dominant peaks will be performed. Accelerometer data will be reviewed, and analysis results and information from other plants will be used, as needed, to evaluate the observations from the SONGS 1 data.
- 2.
If the outer or "alert curve" levels are reached at any frequency, the NRC will be informed within one day as indicated in License Condition 3.M, Section 4.a. Studies will be performed to establish the causes of the increase or decrease in level.
Within 14 days, the conditions will be evaluated and a report provided to the NRC documenting future plans and actions per License Condition 3.M, Section 4.b.
Loose Parts Monitoring Methodology The second method of detecting potential thermal shield problems is by use of the Loose Parts Monitoring system.
This system uses four (4) accelerometers for loose parts monitoring. The accelerometers are mounted at 90* intervals around the circumference of the reactor vessel upper flange at approximately the same elevation as the core barrel seating flange.
License Condition 3.M, Section 3.1 requires that a minimum of two (2) accelerometers be operable in Mode 1.
Spectral characteristics at normal background levels were recorded to establish a baseline for later comparison.
This was done at 91% power and will be used as a basis for comparison with subsequent spectral plots when the situation warrants, i.e., when alarm limits are exceeded or abnormal noise is
noted. RMS voltage values are recorded for a period of five minutes twice a day in accordance with License Condition 3.m, Section 3.1. The values are trended to detect any changes in noise level and are analyzed for possible loose parts activity.
The trends are also used to determine if there is any accelerometer signal degradation.
The recording of all loose parts monitoring activity, including RMS values and audio background noise levels, is performed twice daily. Once a week the cognizant engineer and a member from the Performance Monitoring Group (PMG) review the data for any anomalies.
Audio recordings have been collected at various power levels and plant conditions to establish a library of audio background noise level baselines.
If anomalies are noted, recordings will be compared to the baseline recordings (see Loose Parts Monitoring System Final Acceptance Criteria below).
Accelerometer Impact Sensitivity Essential to early loose part detection is system sensitivity. Regulatory Guide 1.133 specifies the minimum sensitivity requirements for a Loose Parts Monitoring System. The requirement is that the system can detect a metallic loose part that impacts with a kinetic energy of 0.5 ft-lbf on the inside surface of the reactor coolant pressure boundary within three feet of a sensor. On May 12, 1991, with the Unit at normal operating temperature and pressure, impact tool tests were conducted at various locations on the vessel and on the reactor coolant piping at distances greater than 30 feet from the sensors.
The results of these tests, as evaluated in Ref. 8, indicate that 0.5 ft-lbf impacts can be detected at distances as great as 30 feet from the sensors.
Impacts likely to result from loose parts activity can, therefore, be detected at the bottom of the vessel.
Loose Parts Monitoring System Final Acceptance Criteria Impact test data collected in Mode 3 (used to determine accelerometer sensitivity) and data collected at 0% and 91% power were sent to Combustion Engineering (CE) for review and evaluation. The data along with historical data obtained during the previous fuel cycle, were used to establish the basis for the following acceptance criteria:
If the ratio of the peak to average RMS value exceeds four times the average RMS ratio during the 5 minute monitoring interval, or in two or more automatically recorded 10 second data sets, the recording system will be set to count peaks which exceed four times the average ratio.
If the average count rate exceeds one per ten minutes over a two week period, the NRC will be informed within one day of the end of the two weeks, as indicated in License Condition 3.M, Section 4.1. Within 14 days the condition will be evaluated and a report provided to the NRC documenting findings, future plans and action per License Condition 3.M, Section 4.b.
Based on the evaluation of data collected during Cycle 11 and earlier historical data, it has been determined that peak RMS signals two times the
standard deviation of the average RMS values are normal activity due to background noise. It is expected that the average RMS levels will be in the range of 0.12 to 0.25 volts, with the acceptance criteria values set in the range of.48 to 1.0 volts.
References
- 1.
F. J. Sweeny, J, Mach-Leuba and C. M. Smith, "Contribution of Fuel Vibrations to Ex-Core Neutron Noise During the First and Second Fuel Cycles of the Sequoia 1 Pressurized Water Reactor," Progress in Nuclear Energy, 1985, Vol. 15, pp. 283 - 290.
- 2.
F. J. Sweeny, "Sensitivity of Detecting In-Core Vibration and Boiling in Pressurized Water Reactors Using Ex-Core Neutron Detectors",
NUREG/CR-2996 (ORNL/TM-8549), July, 1984.
- 3.
C. Puyal, C. Vincent, C. Meuwisse and A. Trenty, "Use of Low Frequency Fluctuations for the Surveillance of Structures, Sensors and Thermo hydraulic Phenomena on the 900 MW and 1300 MW Reactors", SMORN V, Munich, Germany, October 1987.
- 4.
G. P. Horne, "Recent Experience in Neutron Noise Analysis at Oconee Nuclear Station", Progress in Nuclear Energy, 1985, Vol. 15, pp.
395-402.
- 5.
N. R. Singleton, et. al., "SONGS 1 Cycle 11 Thermal Shield Monitoring Surveillance Criteria", Westinghouse Electric Corporation Document WCAP 13058, dated August, 1991. (Proprietary)
- 6.
J. E. Goossen, et. al., "Additional Engineering Evaluations for the SONGS 1 Thermal Shield Lower Support Replacement Design", Westinghouse Electric Corporation Document WCAP 12981, dated June, 1991.
(Proprietary)
- 7.
J. E. Goossen, A. J. Kuenzel and C. Yu, "As-Built Report for the SONGS 1 Thermal Shield Support Replacement Design", Westinghouse Electric Corporation Document WCAP 13047, dated August, 1991.
(Proprietary)
CRITERIA.NEW
Figure 1 Nuclear Noise Criteria Curve Detectors 5a and 7a 1 C-OO IE07./rti boundaries.
1 0
.~ ~ ~.-.......................................................
.~
...... S ld ln sd n t lr e e rt ra.......
L...
1 7 :..................
o
.E-.:.....
I..........
~~~~.
S-....
lE-07 1E -08:..........
0 2
4 6
8 10 12 14 16 Frequency (Hz)
Figure 2 Nuclear Noise Criteria Curve Detectors 6a and 8a 1E-6 E-06 Soi ie eoeaetlvlciei bo u n d aries...........
- *I 0..-.
\\..
i.........
?.
C..w..................
crie.......aies I.....................................
5..
14.........
.ElO
- 1 1 E -0 7 =.6 E-08=
(Hz)
E -0 9 -.........................
1.........
S o li I...nes...
denote al r level criteria..............
b o u n d a rie s........
Dashed........
lines de oeevlaio.ee
........... ~ ~
~
~
~
~
~
~
~
~
~
~
crte i boundaries..........
........................................I.......
E
-1 O=........
1 C -1 1..........
1 0................
1......................
2........ 1 4.
1...........
6.......
Fre q u e n c y
( H z )..............
t
Figure 3 Nuclear Noise Criteria Curve Detectors 5b and 7b 1E-06......................
.. Solid ines denote alert level criteria 0
.~
.)...................1 on a is CO)
Dashed Ilnes denote evaluation level criteria boundaries.
0 1E-09 V..
~
\\-\\
1E-10 o 2
4 6
01 41 Frequency (Hz)
Figure 4 Nuclear Noise Criteria Curve Detectors 6b and 8b 1
............................................... S i lines........ denote alr level.. criteria..................
Dashed lines denote evaluation level criteria boundaries.
1.....
7...................................
0.....
....................... 1.. E-8...........
.a
..........N........
S~
~ ~ ~~~~~............
.4.
.o......;.............
.\\.....
\\....
~-...........
C/)
\\........
\\....
0-
.~............Z
........ I.......................
0..
S-...-.
~
o 2.4.6.8.10.12.14 1
.~ ~
~
~
r q e c (H z)
Fractional Change per EFPD 0
0 0
0 0
0 b
b0 b
0 0
b0 o
0 0
0 0
0 0
0 0-AL o
- 0)
IQ 01 (C
zCD C
CDC C